Prescreening for oncology clinical trials has become increasingly slow and costly. This paper presents a description of an AI-powered prescreening tool that uses natural language processing (NLP) to identify crucial clinical information from de-identified patient records.